U.S. patent application number 13/521086 was filed with the patent office on 2012-11-22 for defect inspection method and device thereof.
Invention is credited to Toshifumi Honda, Yukihiro Shibata, Atsushi Taniguchi, Taketo Ueno.
Application Number | 20120296576 13/521086 |
Document ID | / |
Family ID | 44367807 |
Filed Date | 2012-11-22 |
United States Patent
Application |
20120296576 |
Kind Code |
A1 |
Shibata; Yukihiro ; et
al. |
November 22, 2012 |
DEFECT INSPECTION METHOD AND DEVICE THEREOF
Abstract
The present invention relates to a defect inspection device
which includes: irradiating means for simultaneously irradiating
different regions on a sample with illumination light under
different optical conditions, the sample being predesigned to
include patterns repeatedly formed thereupon, wherein the patterns
are to be formed in the same shape; detection means for detecting,
for each of the different regions, a beam of light reflected from
each region irradiated with the illumination light; defect
candidate extraction means for extracting defect candidates under
the different optical conditions for each of the different regions,
by processing detection signals corresponding to the reflected
light which is detected; defect extraction means for extracting
defects by integrating the defect candidates extracted under the
different optical conditions; and defect classifying means for
calculating feature quantities of the extracted defects and
classifies the defects according to the calculated feature
quantities.
Inventors: |
Shibata; Yukihiro;
(Fujisawa, JP) ; Honda; Toshifumi; (Yokohama,
JP) ; Ueno; Taketo; (Kawasaki, JP) ;
Taniguchi; Atsushi; (Fujisawa, JP) |
Family ID: |
44367807 |
Appl. No.: |
13/521086 |
Filed: |
February 9, 2011 |
PCT Filed: |
February 9, 2011 |
PCT NO: |
PCT/JP2011/052787 |
371 Date: |
July 30, 2012 |
Current U.S.
Class: |
702/40 |
Current CPC
Class: |
G01B 11/303 20130101;
H01L 21/67288 20130101; G01N 21/956 20130101 |
Class at
Publication: |
702/40 |
International
Class: |
G01N 21/55 20060101
G01N021/55; G06F 19/00 20110101 G06F019/00 |
Foreign Application Data
Date |
Code |
Application Number |
Feb 10, 2010 |
JP |
2010-027702 |
Claims
1. A defect inspection device comprising: irradiating unit which
simultaneously irradiates different regions on a sample with
illumination light under different optical conditions, on the
surface of the sample patterns are repeatedly formed which are
essentially having the same shape in design; detection unit which
detects, for each of the different regions, light reflected from
each region simultaneously irradiated with the illumination light
under the different optical conditions by the irradiating unit;
defect candidate extraction unit which extracts defect candidates
under the different optical conditions for each of the different
regions, by processing detection signals corresponding to the
reflected light detected for each different region by the detection
unit; defect extraction unit which extracts defects by integrating
the defect candidates extracted under the different optical
conditions for each different region by the defect candidate
extraction unit; and defect classifying unit which classifies
feature quantities of the defects extracted by the defect
extraction unit and classifying the defects according to the
calculated feature quantities.
2. The defect inspection device according to claim 1, wherein: the
irradiating unit irradiates different linear regions on the sample,
with beams of illumination light simultaneously and linearly formed
under different optical conditions.
3. The defect inspection device according to claim 1, wherein: the
irradiating unit simultaneously irradiates the sample under optical
conditions of different incident angles of the illumination light
upon the sample.
4. A defect inspection device comprising: a first irradiating unit
which irradiates, with first illumination light via an objective
lens, a surface of a sample patterns are repeatedly formed which
are essentially the same shape in design; a second irradiating unit
which irradiates the sample with second illumination light from an
exterior of the objective lens; reflected-light detection unit
which obtains a first detection signal by detecting only reflected
light which has passed through the objective lens and does not
include regularly reflected light, of all reflected light from a
region irradiated with the illumination light by the first
irradiating means, the reflected-light detection unit further
obtains a second detection signal by detecting only reflected light
which has passed through the objective lens and does not include
regularly reflected light, of all reflected light from a region
irradiated with the illumination light by the second irradiating
unit; and signal-processing unit which detects defects on the
sample by processing the first detection signal and second
detection signal obtained by the reflected-light detection
unit.
5. The defect inspection device according to claim 4, wherein: the
first irradiating unit emits the first illumination light via the
objective lens from a direction inclined with respect to a
normal-line direction of the surface of the sample.
6. A defect inspection device comprising: illumination light
irradiator which irradiates a surface of a sample with illumination
light, on the surface of the sample patterns are repeatedly formed
which are essentially having the same shape in design;
reflected-light detector which simultaneously detects, under
different detection conditions, a beam of light reflected from a
region illuminated with the illumination light by the illumination
light irradiator; defect candidate extractor which extracts defect
candidates for each of the different detection conditions by
processing detection signals obtained during the simultaneous
detection of the reflected light under the different detection
conditions by the reflected-light detector; defect extractor which
extracts defects on the sample by integrating the defect candidates
extracted for each of the different detection conditions by the
defect candidate extractor; and defect classifier which classifies
feature quantities of the defects extracted by the defect extractor
and classifies the defects according to the calculated feature
quantities.
7. The defect inspection device according to claim 6, wherein: the
detection conditions under which the reflected-light detector
detects reflected light are polarizing conditions of the reflected
light.
8. A defect inspection method comprising: simultaneously
irradiating different regions on a sample with illumination light
under different optical conditions, on the surface of the sample
patterns are repeatedly formed which are essentially having the
same shape in design; detecting, for each of the different regions,
a beam of light reflected from each region simultaneously
irradiated with the illumination light under the different optical
conditions; extracting defect candidates under the different
optical conditions for each of the different regions, by processing
detection signals corresponding to the beams of reflected light
detected for each different region; extracting defects by
integrating the defect candidates extracted under the different
optical conditions for each different region; and calculating
feature quantities of the detected defects and classifying the
defects according to the calculated feature quantities.
9. The defect inspection method according to claim 8, further
comprising: irradiating different linear regions on the sample,
with beams of illumination light simultaneously and linearly formed
under different optical conditions.
10. The defect inspection method according to claim 8, wherein: the
different optical conditions differ from each other in an incident
angle of the illumination light upon the sample.
11. A defect inspection method comprising: irradiating, with first
illumination light via an objective lens, a surface of a sample
patterns are repeatedly formed which are essentially the same shape
in design; obtaining a first detection signal by detecting only
reflected light which has passed through the objective lens and
does not include regularly reflected light, of all reflected light
from a region irradiated with the first illumination light;
irradiating the surface of the sample with second illumination
light from an exterior of the objective lens; obtaining a second
detection signal by detecting only reflected light which has passed
through the objective lens and does not include regularly reflected
light, of all reflected light from a region irradiated with the
second illumination light; and detecting defects on the sample by
processing the first detection signal and the second detection
signal.
12. The defect inspection method according to claim 11, further
comprising: irradiating the sample with the first illumination
light via the objective lens from a direction inclined with respect
to a normal-line direction of the surface of the sample.
13. A defect inspection method comprising: irradiating, with
illumination light, a surface of a sample patterns are repeatedly
formed which are essentially the same shape in design;
simultaneously detecting, under different detection conditions,
light reflected from a region illuminated with the illumination
light; extracting defect candidates for each of the different
detection conditions by processing detection signals for each of
the different detection conditions, the detection signals being
obtained during the simultaneous detection of the light reflections
under the different detection conditions; extracting defects on the
sample by integrating the defect candidates extracted for each of
the different detection conditions; and calculating feature
quantities of the extracted defects and classifying the defects
according to the calculated feature quantities.
14. The defect inspection device according to claim 13, wherein:
the different detection conditions are polarizing conditions of the
reflected light.
Description
TECHNICAL FIELD
[0001] The present invention relates to methods for inspecting
defects, foreign matter, and other unwanted substances present on
surfaces of microstructured patterns formed on samples through
thin-film processes represented by semiconductor-manufacturing
processes and flat-panel display manufacturing processes. The
invention also relates to devices used in the above defect
inspection methods.
BACKGROUND ART
[0002] Examples of prior-art semiconductor inspection device
configurations include the configuration described in
JP-T-2005-517906. This prior-art inspection device has a dark-field
detection optical system that illuminates a wafer surface with
laser light obliquely and detects the light scattered from the
wafer surface. The optical system employs off-axis illumination in
which the laser light is emitted from the outside of an objective
lens. During off-axis illumination, if the optical system includes
vertical detection optics having an optical axis parallel to a
normal line relative to the surface of the wafer, the system needs
to have a low elevation angle for illumination due to mechanical
restrictions associated with disposition of the vertical detection
optics.
[0003] On the other hand, JP-A-2000-2664 describes through-the-lens
(TTL) illumination in which light is emitted through an objective
lens having an optical axis vertical to a wafer. In the TTL
illumination, the wafer can be illuminated substantially
vertically.
PRIOR ART LITERATURES
Patent Document
[0004] Patent Document 1: JP-T-2005-517906 [0005] Patent Document
2: JP-A-2000-2664
SUMMARY OF THE INVENTION
Problem to be Solved by the invention
[0006] Various patterns are formed on semiconductor wafers. The
kinds of defects also vary from cause to cause. Patterns
represented by DRAMs (Dynamic Random Access Memories) and the like,
are periodically wired patterns, and examples of serious defects
influential upon a semiconductor device production yield include
defects that can cause pattern short-circuiting, as well as
scratches. In particular, short-circuiting defects at the groove
bottom of an etched pattern prevent illumination light from
reaching the groove bottom in such cases as too small an
illumination elevation angle or too narrow an interconnecting
pitch, and thereby reduce the amount of light scattered from the
short-circuiting defects at the groove bottom. This is most likely
to result in the defects being overlooked.
[0007] In addition, for a recessed defect such as a scratch, since
increasing the illumination elevation angle provides a larger
cross-sectional area for scattering, higher-elevation illumination
enables a larger amount of light to be scattered from the scratch.
For this reason, higher-elevation illumination is advantageous for
detecting any groove bottom short-circuiting defects and scratches
present on etched pattern surfaces.
[0008] The off-axis illumination described in prior-art Patent
Document 1, however, has a problem in that the presence of the
restrictions which make it absolutely necessary for the optical
system to have a low elevation angle for illumination can become an
obstruction to implementing high elevation angle illumination
advantageous for detecting the above defects.
[0009] Additionally, although the TTL illumination described in
prior-art Patent Document 2 enables high elevation angle
illumination, the TTL illumination has a problem in that during
dark-field detection that uses a spatial filter to block regularly
reflected light, stray light such as lens-reflected light reaches
an image surface and is thus likely to affect inspection
sensitivity.
[0010] The present invention provides a dark-field defect detection
method and related device based on high elevation angle
illumination, the detection method and related device being adapted
to resolve the above problems and thus during detection of any
groove bottom short-circuiting defects and scratches present on
etched pattern surfaces, prevent lens-reflected light and other
stray light from reaching an image surface and affecting inspection
sensitivity.
Means for Solving the Problems
[0011] In order to solve the above problems, a defect inspection
method and related device according to a first aspect of the
present invention is designed to: simultaneously irradiate
different regions on a sample with illumination light under
different optical conditions, on the surface of the sample patterns
are repeatedly formed which are essentially having the same shape
in design; detect, for each of the different regions, a beam of
light reflected from each region simultaneously irradiated with the
illumination light under the different optical conditions; extract
defect candidates under the different optical conditions for each
of the different regions, by processing detection signals
corresponding to the beams of reflected light detected for each
different region; extract defects by integrating the defect
candidates extracted under the different optical conditions for
each different region; and calculate feature quantities of the
detected defects and classify the defects according to the
calculated feature quantities.
[0012] Additionally, in order to solve the above problems, a defect
inspection method and related device according to a second aspect
of the present invention is designed to: irradiate, with first
illumination light via an objective lens, a surface of a sample
patterns are repeatedly formed which are essentially the same shape
in design; obtain a first detection signal by detecting only
reflected light which has passed through the objective lens and
does not include regularly reflected light, of all reflected light
from a region irradiated with the first illumination light;
irradiate a surface of the sample with second illumination light
from outside of the objective lens; obtain a second detection
signal by detecting only reflected light which has passed through
the objective lens and does not include regularly reflected light,
of all reflected light from a region irradiated with the second
illumination light; and detect defects on the sample by processing
the first detection signal and the second detection signal.
[0013] Furthermore, in order to solve the above problems, a defect
inspection method and related device according to a third aspect of
the present invention is designed to: irradiate, with illumination
light, a surface of a sample patterns are repeatedly formed which
are essentially the same shape in design; simultaneously detect,
under different detection conditions, a beam of light reflected
from a region irradiated with the illumination light; extract
defect candidates for each of the different detection conditions by
processing detection signals for each of the different detection
conditions, the detection signals being obtained during the
simultaneous detection of the light reflections under the different
detection conditions; extract defects on the sample by integrating
the defect candidates extracted for each of the different optical
conditions; and calculate feature quantities of the extracted
defects and classify the defects according to the calculated
feature quantities.
Effect of the Invention
[0014] In accordance with the present invention, images
advantageous for more highly sensitive inspection of more easily
identifiable defects can be obtained by efficiently detecting high
elevation angle illumination light scattered from varieties of
defects present on a wafer, such as groove bottom defects and
scratches.
BRIEF DESCRIPTION OF THE DRAWINGS
[0015] FIG. 1 is a block diagram showing a schematic configuration
of optical systems based on off-axis illumination;
[0016] FIG. 2 is a block diagram showing a configuration of an
image-processing unit in a first embodiment;
[0017] FIG. 3 is a block diagram showing a schematic configuration
of TTL-illumination-based optical systems in a second
embodiment;
[0018] FIG. 4 is a block diagram showing a schematic configuration
of a TTL-illumination-based vertical illumination optical system in
the second embodiment;
[0019] FIG. 5 is a block diagram showing a configuration of an
image-processing unit in the second embodiment;
[0020] FIG. 6 is a block diagram showing a schematic configuration
of a TTL-illumination-based oblique illumination optical system in
a third embodiment, the optical system being constructed so as to
prevent regularly reflected light from entering an objective
lens;
[0021] FIG. 7 is a block diagram showing a schematic configuration
of a TTL-illumination-based oblique illumination optical system in
a modification of the third embodiment, the optical system being
constructed to cause regularly reflected light to enter an
objective lens;
[0022] FIG. 8 is a block diagram showing a schematic configuration
of a TTL-illumination-based oblique illumination optical system in
another modification of the third embodiment, the optical system
being constructed to irradiate a checkered pattern with
illumination light and detect the light reflected from the
pattern;
[0023] FIG. 9 is a block diagram showing a simultaneous
illumination optical system configuration that uses off-axis
illumination and TTL illumination in a fourth embodiment;
[0024] FIG. 10 is a block diagram showing a schematic configuration
of a TTL-illumination-based vertical illumination optical system in
a fifth embodiment, the optical system being constructed to detect
a plurality of images different from one another in detection
conditions;
[0025] FIG. 11 is a plan view of a polarizing filter in the fifth
embodiment;
[0026] FIG. 12 is a block diagram showing a configuration of an
image-processing unit in the fifth embodiment;
[0027] FIG. 13 is an azimuthal diagram of polarization that shows
four kinds of polarized-light transmission axes in the fifth
embodiment; and
[0028] FIG. 14 is a scatter diagram of defects with an azimuth
angle plotted on a vertical axis and ellipticity on a horizontal
axis.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0029] Hereunder, embodiments of the present invention will be
described using the accompanying drawings.
First Embodiment
[0030] A configuration of a semiconductor wafer defect inspection
device according to the present invention is shown in FIG. 1. A
wafer 1 is mounted on a stage 3 and then a scanning direction of
the stage and patterns formed on the wafer 1 are made parallel in a
horizontal direction (this operation is referred to as O-alignment
of the wafer 1), whereby coordinates are matched between a
reference position of the wafer 1 and an X-Y coordinate system of
the stage 3. The stage 3 can be moved independently in an X-Y plane
and in a Z-direction perpendicular thereto, and is further
constructed to be rotatable about an axis of a Z-direction. An
illumination optical system 4, disposed obliquely to the wafer 1,
linearly illuminates the surface of the wafer 1 with linear beams
of illumination light, 29 and 30.
[0031] The illumination optical system 4 uses either a laser or a
lamp, as its light source 5. In case a laser is used as the light
source 5, the laser will be either a YAG second-harmonic 532-nm
laser, third-harmonic 355-nm laser, fourth-harmonic 266-nm laser,
199-nm laser, or 193-nm laser, or some other laser that emits
deep-ultraviolet (DW) light. A multispectral laser that oscillates
at a plurality of wavelengths will be a further candidate. In case
a lamp is used as the light source 5, candidates will be mercury
lamps or mercury-xenon lamps that emit d-line (588 nm) light,
e-line (546 nm) light, g-line (436 nm) light, h-line (405 nm)
light, and/or i-line (365 nm) light. The embodiment shown in FIG. 1
uses a laser as the light source 5.
[0032] Laser light which is emitted from the laser light source 5
enters an electro-optic element 7 (such as LiNb03 or PLZT [short
for (Pb, La)(Zr, Ti)03]) that can electrically control polarization
in a predetermined direction. This electro-optic element may be
replaced by a magneto-optic element formed from a garnet film, for
example. Upon the polarizing direction being controlled, the amount
of light passing through a polarizing beam splitter (PBS) 8 is
reduced to a predetermined level and after the light has entered a
beam splitter 9, part of the light reflects therefrom and the rest
passes therethrough.
[0033] The light that has reflected from the beam splitter 9 is
expanded in beam diameter by a beam expander 11, and then at a
cylindrical lens 23, the expanded beam of light is converged in one
direction and components at right angles to this converging
direction are linearly shaped as parallel beams. This linearly
shaped illumination light 25 reflects from a mirror 24 and enters a
linear region 29 on the wafer 1 at incident angles of 45.degree. to
90.degree. to provide high elevation angle illumination, by
utilizing a space created above the wafer 1 as a result of setting
up an objective lens 31 slantwise with respect to a normal
direction of the wafer 1.
[0034] On the other hand, the light that has passed through the
beam splitter 9 enters a beam expander 10 and is expanded in beam
diameter thereby. This expanded beam of light has its optical path
oriented towards the wafer 1 by mirrors 12 and 13, and the beam is
polarized to a predetermined state by a half wavelength plate 15
and a quarter wavelength plate 17, both rotatable. The polarization
here is, for example, either S-polarization, P-polarization, linear
polarization that generates oscillations at an angle somewhere in
between S-polarization and P-polarization angles, or clockwise or
counterclockwise elliptical polarization, with respect to the wafer
1. The illumination light 22, after passing through the half
wavelength plate 15 and the quarter wavelength plate 17, is
linearly shaped by a cylindrical lens 20 to conduct low elevation
angle illumination of a linear region 30 (a thin line narrow in a
direction of X and long in a direction of Y) on the wafer 1 at
incident angles of 2.degree. to 45.degree..
[0035] After the low-angle illumination (low elevation angle
illumination) of the linear region 30 on the wafer 1 by the
illumination light 22, of all components of the light scattered
from the illuminated linear region 30, only those propagating
within a numerical aperture (NA) of an objective lens 31, which is
installed by inclining with respect to the normal line of the wafer
1, enter a detection optical system 50, thus forming an image on a
detection surface (not shown) of an image sensor 90b via an
image-forming lens 45.
[0036] Meanwhile, after the high elevation angle illumination of
the linear region 29 on the wafer 1 by the illumination light 25,
of all components of the light scattered from the linear region 29
to which the objective lens 31 is closer to the wafer 1 than the
linear region 30 by the inclination of the objective lens 31, only
components that are to have an elevation angle lower than the NA of
the objective lens 31 are reflected within the NA of the objective
lens 31 by a mirror 32. A distance from the linear region 29 via
the mirror 32 to the objective lens 31 is made equal to a working
distance (WD). Thus, of the light that has been scattered from the
linear region 29 of the wafer 1, the components that have reflected
from the mirror 32 and entered the objective lens 31 form an image
on a detection surface (not shown) of an image sensor 90a via the
image-forming lens 45.
[0037] This configuration of the wafer defect inspection device
enables it to simultaneously detect the two kinds of dark-field
images (formed by high elevation angle illumination and low
elevation angle illumination) that differ in illumination elevation
angle at spatially separate positions on the wafer 1. The spatially
separate positions refer to the wafer linear region 29 irradiated
with the illumination light 25, and the wafer linear region 30
irradiated with the illumination light 22. The image formed by the
scattered light caused by the low elevation angle illumination
light and detected by the image sensor 90a, and the image formed by
the scattered light caused by the high elevation angle illumination
light and detected by the image sensor 90b are input to an
image-processing unit 100. Then, in the image-processing unit 100,
input two images are compared with an image of the same pattern on
design (e.g., an image of an adjacent die). Thus, defects are
detected.
[0038] Defect determination and defect classification, both based
on features and characteristics of images that differ in detection
elevation angle, are also possible by position-matching the images
of the same coordinates on the wafer 1 that have been detected by
the image sensors 90a and 90b respectively. The image sensors 90a,
90b are, for example, charge-coupled device (CCD) or complementary
metal-oxide semiconductor (CMOS) sensors, or a TDI (Time Delay
Integration) operation type based on these sensors can be used as
an alternative.
[0039] FIG. 2 shows a flow of processing by the image-processing
unit 100 which conducts defect determination by processing the
images detected by the image sensors 90a and 90b.
[0040] The image detected by the image sensor 90a is subjected to
conversion of brightness, such as gamma(.gamma.)-correction, in a
gradation conversion unit 101a. The output image A from the
gradation conversion unit 101a is divided in two and one of the two
images A is sent to a position-matching unit 105a, and the other is
sent to a memory 103a. The position-matching unit 105a receives,
from the memory 103a, an image A' of a pattern which is essentially
the same pattern in design (e.g., image of an adjacent die) that
has already been detected by the image sensor 90a and stored into
the memory 103a, and then matches positions of the images A and
A'.
[0041] A comparator 107a creates a differential image B from the
position-matched images A and A' by comparing the images A and A'
with a threshold level that is either a previously set value or a
value previously calculated from the detected images, and then
calculates feature quantities of the differential image B as
results of the comparisons. Next, a defect-determining unit 115
uses the feature quantities (such as a maximum contrast level and
area) of the differential image B to determine whether the image
contains defects.
[0042] Substantially the same processes as those described above
are also conducted upon the image detected by the image sensor 90b.
That is to say, the image detected by the image sensor 90b is
subjected to the conversion of brightness, such as
.gamma.-correction, in a gradation conversion unit 101b. The output
image C from the gradation conversion unit 101b is divided in two
and one of the two images C is sent to a position-matching unit
105b, and the other is sent to a memory 103b. The position-matching
unit 105b receives, from the memory 103b, an image C' of a pattern
which is essentially the same pattern in design (e.g., image of an
adjacent die) that has already been detected by the image sensor
90b and stored into the memory 103b, and then matches positions of
the images C and C'.
[0043] A comparator 107b creates a differential image D from the
position-matched images C and C' by comparing the images C and C'
with a threshold level that is a previously set value or a value
calculated from the detected images, and calculates feature
quantities of the differential image D as results of the
comparisons. Next, the defect-determining unit 115 uses the feature
quantities (such as a maximum contrast level and area) of the
differential image D to determine whether the image contains
defects.
[0044] Additionally, the image comparison results by the
comparators 107a, 107b are sent to a position-matching unit 111, in
which the differential images B and D different in illumination
elevation angle are then further matched in position. A
differential image comparator 112 compares feature quantities of
these differential images detected under the different optical
conditions, and then the feature quantities are sent to the
defect-determining unit 115 for defect determination. In this way,
the defect-determining unit 115 uses the three kinds of feature
quantities to conduct determinations. If any one of the three sets
of determination results indicates that the corresponding image is
defective, the feature quantities of the remaining two kinds of
images are sent with the particular image to a classification unit
117 as information. The classification unit 117 classifies detected
defects by kinds (e.g., foreign matter, etching residues, or
scratches) or as dummy defects (such as non-uniformity in
brightness of an oxide film, roughness of the pattern, grains, or
other factors not critical or fatal to the semiconductor device).
Coordinates, classification results, feature quantities, and others
of the defects are sent to an operating unit 110, such that a user
of the inspection device can display and output defect information
data, a map of the defects on the wafer, and other defect
information.
[0045] The coordinates, dimensions, and brightness of the detected
defects, the features and characteristics of each defect that the
differences in detection elevation angle will elucidate, and other
defect information are sent to the operating unit 110, such that
the user of the inspection device can check the display which
outputs the defect information data, the on-wafer defect map, and
other defect information.
[0046] The operating unit 110 also has a function for assigning
operational instructions relating to the inspection device, and
controls operation of a stage 3 and optical parts from a mechanism
control unit 120 by giving operational instructions to the
mechanism control unit 120. The detection optical system may have a
spatial light modulator (not shown) on a Fourier transform plane of
the wafer 1. The spatial light modulator displaced in that case may
be a micro-shutter array that utilizes disposition of a metallic
light-blocking rod or electro-optic effects of a birefringent
element (such as LiNb03 or PLZT [short for (Pb, La)(Zr, Ti)03]).
Alternatively, the spatial light modulator may be a liquid-crystal
filter or a one-dimensional and two-dimensional array of filters
that uses MEMS (Micro-Electro Mechanical Systems).
[0047] These devices can switch light transmitting/blocking rapidly
by electrical control, and thus during inspection, enables changing
to an appropriate filtering pattern according to a particular pitch
and shape of a pattern 2 present on the wafer 1. In addition, to
match height of a surface layer of the wafer 1 to a focal position
of an objective lens 31, it is necessary to control wafer height by
detecting the height of the wafer 1 and controlling a Z-axial
position of the stage 3.
[0048] There is an optical lever method as an example of a method
useable for wafer height detection. Although not shown, a height
detection illumination system that obliquely illuminates a wafer 1
with slit light, and a height detection system that calculates
wafer height by detecting the slit light reflected from the wafer 1
are arranged in the optical lever method. A difference between
height of the wafer 1 and a focal position of an objective lens 31
is calculated and in case of a defocusing tolerance being
overstepped, the mechanism control unit 120 instructs the stage 3
to adjust the height of the wafer 1 to the focal position. In the
basic configuration described above, the inspection device
continuously detects images while moving the stage 3 in one
direction for constant speed scanning in the X-Y plane, and detects
defects using the images.
[0049] Although the illumination regions 29 and 30 on the wafer 1
have been taken as an example of linear regions subjected to high
elevation angle illumination and low elevation angle illumination,
respectively, in the present embodiment, the high elevation angle
illumination and low elevation angle illumination regions may be
interchanged or both regions may be illuminated from the same
elevation angle. Furthermore, polarization, illumination azimuths
(directions of viewing angles relative to the X-axis when the wafer
1 is seen in plan view from above), illumination wavelengths, and
other optical conditions may be changed for both of the
illumination regions 29 and 30.
[0050] Still furthermore, there is a case having a characteristic
in scattering directions, depending on defect sizes and kinds
(short-circuiting defects, scratches, foreign substances, and so
on). Simultaneously detecting low elevation angle illumination
light and high elevation angle illumination light, therefore,
enhances a possibility that the light scattered from various
defects will be detected, and becomes advantageous for improving a
capture ratio of defects.
Second Embodiment
[0051] Next, a second embodiment will be described. In the present
embodiment, the detection optical system 50 in the configuration of
FIG. 1 is disposed in a direction perpendicular to the wafer 1,
high elevation angle illumination that uses illumination light 25
is made common in specifications to an optical axis of the
detection optical system 50, and the same regions on the wafer are
sequentially illuminated by changing optical paths.
[0052] FIG. 3 shows a configuration of a through-the-lens (TTL)
type of optical system for conducting high elevation angle
illumination via an objective lens 331. A laser light source is
shown as 5, a beam expander as 311, an optical-path switching
mirror as 312, a beam splitter as 327, a douser as 333, an
objective lens 331, and a lens as 375. A spatial filter with a
light-blocking element 381 is shown as 380, and an image-forming
lens is shown as 385. Additionally, a mirror is shown as 313, a
rotatable half wavelength plate as 315, a rotatable quarter
wavelength plate as 317, and a cylindrical lens as 3320.
[0053] In this configuration, light which is emitted from the laser
light source 5 enters the beam expander 311, then after being
expanded in beam diameter thereby, travels straight forward without
the optical path switching mirror 312 acting, and enters the beam
splitter 327 disposed on the optical axis of the detection optical
system. After being reflected by the beam splitter 327, the light
further enters an opening device such as a linear opening for
confocal detection or a pinhole array. The light that has passed
through the opening device 333 provides the wafer 1 with high
elevation angle illumination via the objective lens 331. Of the
light scattered from the wafer 1, light that propagates into an NA
of the objective lens 331 is captured by the objective lens 331 and
once again passes through the opening device 333. The light, after
further passing through the beam splitter 327 as well, reaches the
lens 375 and forms a Fourier transform image of the wafer 1 on a
surface having the spatial filter 380 disposed thereon. The spatial
filter 380 includes the light-blocking element 381 that blocks
regularly reflected light (0th-order light) and desired frequency
components, and the light that has passed through regions other
than the light-blocking element 381 forms the wafer image via the
image-forming lens 385, on a surface having the image sensor 390
disposed thereon.
[0054] In this detection confocal optical system, since the opening
device 333 blocks much of the stray light even if the illumination
light reflects from the surface of the objective lens 331 and goes
astray, a detection rate thereof by the image sensor 390
significantly decreases. This makes it possible to suppress stray
light that has been a bottleneck in the blocking of regularly
reflected light in the prior-art in combination of TTL illumination
and spatial filters.
[0055] If the defects to be inspected are convexities such as
foreign substances, using low elevation angle illumination, rather
than increasing the NA of the objective lens 31 in combination with
a high elevation angle illumination, may provide images
advantageous for highly sensitive inspection. In such a case, laser
light that has been emitted from the laser light source 5 is
reflected downward, as shown in FIG. 3, by using the optical-path
switching mirror 312. The thus-reflected light enters a mirror 322
and after passing through the half wavelength plate 315 and the
quarter wavelength plate 317 as low elevation angle illumination
light 322. After adjustment of this light in polarization state,
the cylindrical lens 320 converges the light in one direction in a
cross-sectional plane perpendicular to the optical axis of the
laser. Components of the laser that are perpendicular to the
converging direction are formed into parallel beams, thus providing
the wafer 1 with off-axis illumination by obliquely illuminating
the laser from the outside of the optical axis of the objective
lens. Even with off-axis illumination, confocal detection is
possible by matching the wafer region illuminated with the off-axis
illumination light 22, nearly to an on-wafer projection image of
the opening portion in the opening device 333. The light scattered
from the wafer 1 illuminated by the off-axis illumination light 22
is captured by the objective lens 31 and then passes through the
opening portion in the opening device 333. Light scattered from a
lower-layer region of the wafer 1, on the other hand, becomes
expanded light at the opened device 333, and much of the light is
blocked by the opening device 333. The light scattered from the
wafer surface-layer and passed through the opening device 333
passes through the beam splitter 327 and enters the lens 375. The
light that has passed through the lens and the spatial filter 380
disposed on a Fourier transform plane forms a scatter image of the
wafer 1 on the image sensor 390 via the lens 385.
[0056] Other beneficial effects of confocal detection are described
below. The layer of the wafer 1 that is to be inspected is usually
a top layer of the wafer surface, so defects present on lower
layers having multilayer interconnects formed thereupon are
generally excluded from detection. This is because, even if these
lower-layer defects are to be detected, since the defects are not
identifiable during reviewing with a scanning electron microscope
(SEM), it cannot be confirmed whether they are defects. During
confocal detection, if the objective lens 331 is properly focused
on the surface layer of the wafer 1 that is to be subjected to
defect detection, the light scattered from defects or the pattern
present on the surface layer will pass through the opening portion
in the opening device 333 and enter the beam splitter 327. Half of
this incident light will pass through the beam splitter 327 first
and then the lens 375, and the light, after further passing through
the spatial filter 380, will enter the image-forming lens 385 and
form each on-wafer defect image on a detection surface (not shown)
of the image sensor 390. In contrast to this, the light scattered
from the lower layers of the wafer 1 will be defocused and expanded
at the opening device 333 and much of this light will be blocked by
the opening device 333. The scattered light from the lower layers
that become pseudo defects will therefore be nearly blocked, which
will in turn be effective for suppressing the detection of the
lower-layer defects.
[0057] Wafer structures and the kinds of defects are various, and
optical conditions for obtaining images advantageous for highly
sensitive inspection differ according to the particular subject of
inspection. At many of semiconductor-manufacturing lines, optical
conditions to be used for inspection are made appropriate for each
manufacturing process. Accordingly, low elevation angle
illumination, not high elevation angle illumination that includes
vertical illumination, may be advantageous for highly sensitive
inspection of some specific kinds of defects.
[0058] For these reasons, a wide range of settable illumination
conditions need to be provided beforehand to respond to such
diversity of wafer structures and defect kinds. FIG. 4 shows an
illumination optical system configuration controllable in
illumination light elevation angle. The configuration shown in FIG.
4 is basically the same as the optical system configuration
described in FIG. 3, but differs in that a beam expander 410, an
opening device 411, and a lens 414 are included. Light which is
emitted from a laser light source 5 is expanded in beam diameter by
the beam expander 410, then after passing through an opening
portion in the opening device 411, passes through the lens 414, and
thus forms an image of the opening device 411 in a pupil position
432 of an objective lens 431. A section including a beam splitter
427 and an opening device 433 for confocal detection is of the same
configuration and function as those described using FIG. 3.
[0059] An incident angle at which the wafer 1 is illuminated is
determined according to the imaging position of an opening device
411 at the pupil of the objective lens 431. For example, if an
image of the objective lens 431 is formed in a central position of
the pupil, the illumination is vertical illumination relative to
the wafer 1. On the other hand, if an image of the opening device
411 is offset with respect to an optical axis, the wafer 1 is
obliquely illuminated within the range of the NA of the objective
lens. For this reason, the opening portion in the opening device
411 is constructed to be changeable (controllable) in diameter,
such that the incident angle at which the wafer 1 is illuminated
can be controlled.
[0060] If the defects to be inspected are convexities such as
foreign substances, using low elevation angle illumination, rather
than increasing the NA of the objective lens 431 in combination
with a high elevation angle illumination, may provide images
advantageous for highly sensitive inspection. In such a case, laser
light that has been emitted from a laser light source 5 is
reflected downward, as shown in FIG. 4, by using an optical-path
switching mirror 412. The thus-reflected light enters a mirror 413
and after passing through the half wavelength plate 415 and the
quarter wavelength plate 417 as low elevation angle illumination
light 422. After adjustment of this light in polarization state, a
cylindrical lens 420 converges the light in one direction in a
cross-sectional plane perpendicular to an optical axis of the
laser. Components of the laser that are perpendicular to the
converging direction are formed into parallel beams and this
illumination light 422 provides the wafer 1 with off-axis
illumination by obliquely illuminating the laser from the outside
of the NA of the objective lens 431. Even with off-axis
illumination, confocal detection is possible by matching the wafer
region illuminated with the off-axis illumination light 422, nearly
to an on-wafer projection image of the opening portion in the
opening device 433.
[0061] The light scattered from the wafer 1 illuminated by the
off-axis illumination light 422 is captured by the objective lens
431 and then passes through the opening portion in the opened
device 433. Light scattered from a lower-layer region of the wafer
1, on the other hand, becomes expanded light at the opened device
433, and much of the light is blocked by the opening device 433.
The light scattered from the wafer surface-layer and passed through
the opening device 433 passes through the beam splitter 427 and
enters the lens 475. The light that has passed through the analyzer
477 and the spatial filter 480 disposed on a Fourier transform
plane forms a scatter image of the wafer 1 on the image sensor 490
via the lens 485.
[0062] Next, a flow of processing conducted by an image-processing
unit 300 which processes the image that has been detected by the
image sensor 390 (490), and determines whether the image contains
defects, is shown in FIG. 5.
[0063] The image A.sub.0 detected during high elevation angle
illumination by the image sensor 390 (490) is subjected to
conversion of brightness, such as .gamma.-correction, in a
gradation conversion unit 501. The output image from the gradation
conversion unit 501 is divided in two and one of two images E is
sent to a position-matching unit 503, and the other (image E') is
sent to a memory 502. The position-matching unit 503 receives, from
the memory 502, an image E' of a pattern which is essentially the
same pattern (e.g., image of an adjacent die) that has already been
detected by the image sensor 390 (490) and stored into the memory
502, and then matches positions of the images E and E'.
[0064] A comparator 504 creates a differential image F from images
E and E', the images of which are obtained from position-matching
results, by comparing the images E and E' with a threshold level
that is either a previously set value or a value previously
calculated from the detected image, and then calculates feature
quantities of the differential image F as results of the
comparisons. Next, a defect-determining unit 507 uses the feature
quantities (such as a maximum contrast level and area) of the
differential image F to determine whether the image contains
defects. The differential image F is also input to a
position-matching unit 505.
[0065] Next, the image G.sub.0 detected during low elevation angle
illumination by the image sensor 390 (490) is subjected to the
conversion of brightness, such as .gamma.-correction, in the
gradation conversion unit 501. The output image from the gradation
conversion unit 501 is divided in two and one of two images G by
the conversion is sent to the position-matching unit 503, and the
other (image G') is sent to the memory 502. The position-matching
unit 503 receives, from the memory 502, an image G' of a pattern
which is essentially the same pattern in design (e.g., image of an
adjacent die) that has already been detected by the image sensor
390 (490) and stored into the memory 502, and then matches
positions of the image G and an image G'.
[0066] The comparator 504 compares the difference image H, the
differential between the position-matched images G and G', with
respect to a threshold level that is either a previously set value
or a value previously calculated from the detected image, and then
calculates feature quantities of the differential image H as
results of the comparison. Next, the defect-determining unit 507
uses the feature quantities (such as a maximum contrast level and
area) of the differential image D to determine whether the image
contains defects. The differential image H itself is input to the
position-matching unit 505.
[0067] The position-matching unit 505 matches positions of the
differential images F and H that are images of the same place on
the wafer 1, but different in illumination elevation angle. A
differential image comparator 506 then compares feature quantities
of these differential images F and H obtained using different
conditions, and sends the feature quantities to the
defect-determining unit 507 for defect determination. The
defect-determining unit 507 conducts determinations using three
kinds of feature quantities. If any one of the three sets of
determination results indicates that the corresponding image is
defective, the feature quantities including that of the remaining
two kinds of images are sent to a classification unit 508. The
classification unit 508 classifies detected defects by kinds (e.g.,
foreign matter, etching residues, or scratches) or as pseudo
defects (such as non-uniformity in brightness of an oxide film,
roughness of the pattern, grains, or other factors not critical or
fatal to the semiconductor device). Coordinates, classification
results, feature quantities, and others of the defects are sent to
an operating unit 310, such that a user of the inspection device
can check the display which outputs defect information data, a map
of the defects on the wafer, and other defect information.
[0068] The coordinates, dimensions, and brightness of the detected
defects, the features and characteristics of each defect that the
differences in detection elevation angle will elucidate, and other
defect information are sent to the operating unit 310, such that
the user of the inspection device can check the display which
outputs the defect information data, the on-wafer defect map, and
other defect information.
Third Embodiment
[0069] The configuration in which the regularly reflected light
from the wafer, by the TTL illumination, is blocked by a spatial
filter, has been described and shown in the second embodiment. In
the configuration of the second embodiment, forward-scattered light
approximate to the regularly reflected light, that is,
low-frequency components are detected. But some specific kinds of
defects under inspection may strongly distribute back-scattered
light (high-frequency components). The following describes a third
embodiment relating to a scheme which uses TTL illumination to
capture the light strongly distributed in back-scattered light.
FIG. 6 shows an inspection device configuration that employs the
scheme.
[0070] The present embodiment also employs high elevation angle
illumination and low elevation angle illumination. low elevation
angle illumination is of the same optical system configuration and
light detection operation as those of the first and second
embodiments, and description of the low elevation angle
illumination optical system is therefore omitted. Referring to FIG.
6, in a high elevation angle illumination optical system that uses
illumination light 626, when laser light is emitted from a laser
light source 5, a laser beam axis control mechanism 606 controls a
position of incidence, as well as an angle range, on a pupil 662 of
an objective lens 631, and the light enters the objective lens 631
of an illumination and detection optical system 600 inclined
relative to a normal line of a wafer 1. Admitting the laser light
into a right side (see FIG. 6) of the inclined pupil 662 of the
objective lens 631 correspondingly increases an incident angle
relative to the wafer 1 and makes regularly reflected light 660
from a linearly illuminated region on the wafer 1 propagate through
the outside of an NA of the objective lens 631 (the linear
illumination uses the same method as that described in the second
embodiment using FIG. 3). The light captured by the objective lens
631 is therefore backscattered light. A region in an optical path
from a confocal detection opening device 633 to an image sensor 690
is substantially the same as that of FIG. 3 in terms of
configuration.
[0071] A technique for detecting forward-scattered light in this
configuration is described below using FIG. 7. A basic
configuration for the detection of forward-scattered light is the
same as the configuration described in FIG. 6, and in the basic
configuration, light that reaches the pupil 662 of the objective
lens 631 is controlled by a laser beam axis control mechanism 606
to reach a portion close to the normal line of the wafer 1. Thus,
the surface of the wafer 1 is illuminated with nearly vertical
light at a higher angle of elevation than in the embodiment
described in FIG. 6, and forward-scattered light that includes
regularly reflected light is detected. The regularly reflected
light is blocked by a spatial filter 680, and scattered light that
has passed through the spatial filter 680 forms an image on the
image sensor 690 via an image-forming lens 685.
[0072] A modification of confocal detection optical system is shown
in FIG. 8. In a basic configuration of the confocal detection
optical system, its optical axis is inclined with respect to a
normal-line direction of a wafer 1, as described in FIG. 6.
Illumination light 826, which is emitted from a laser light source
(not shown), enters an opening device 840a having a
checker-patterned opening portion for confocal detection. The
opening device 840a has a light-blocking element 842 in addition to
the opening portion 841. An image of the opening device 840a is
projected onto the wafer 1 via an image-forming lens 834 and an
objective lens 831. Of the light that has been scattered from the
wafer 1 illuminated with the illumination light 826, the light
incident in the objective lens 831 passes through an illumination
system/detection system branching mirror 827 via the image-forming
lens 834 and forms an image on a detection surface.
[0073] An opening device 840b of a checkered pattern is placed on
the image-forming plane. The opening portion 841 of the opening
device 840b is disposed in conjugate relationship with respect to
the image of the opening portion disposed in the illumination
system. Light that has passed through the opening device 840b is
branched by a polarized-beam splitter 844, and scattered light of
specific frequencies is blocked by spatial filters 881c, 881d
arranged in respective optical paths. Lights that have passed
through the spatial filters either 881c or 881d enter opening
devices 843c, 843d respectively for confocal detection, and only
lights that have passed through openings in the opened devices
either 843c or 843d enter image sensors either 890c or 890d.
[0074] An example of detecting images of the same space using two
kinds of detection conditions (analyzing conditions and spatial
filtering conditions) has been described and shown in the present
embodiment. Thus, if a plurality of kinds of defects to be detected
are present on the same wafer, detection conditions for exposing
each defect can be assigned and detection sensitivity can be
improved. While the example of using two systems to assign
detection conditions in the present embodiment has been described,
application to an example of using more detectors is also highly
probable.
Fourth Embodiment
[0075] A fourth embodiment that relates to an optical system
configuration consisting of a combination of the optical systems
described in the first and third embodiments is described below
using FIG. 9. The fourth embodiment employs two optical systems.
One system uses high elevation angle TTL illumination light 926
with a laser light source 5, and the other system uses low
elevation angle off-axis illumination light 22. The high elevation
angle TTL illumination light linearly illuminates a region 930 on a
wafer 1 (the illumination of the linear region 930 on the wafer 1
uses the same method as that described per FIG. 6). The off-axis
illumination light 22 linearly illuminates a region 929 (the
illumination of the linear region 29 on the wafer 1 uses the same
method as those described in the first and second embodiments, so
description of the off-axis illumination is omitted). A detection
system has an axis inclined with respect to a normal line of the
wafer 1, and of all light scattered from the regions 929 and 930,
only the light scattered towards an objective lens 931 is captured
by the objective lens 931.
[0076] Here, the light scattered at the region 929 present in the
direction that the detection system is inclined with respect to the
wafer 1 reaches a mirror 932 and works so that the light scattered
at a lower-elevation side than an NA of the objective lens 931 will
then propagate into the NA thereof. A distance from the
illumination region 929 to the objective lens 931, and a distance
from the illumination region 930 to the objective lens 931 are
matched to a working distance (WD), whereby scatter images of the
illumination regions can be formed in parallel on an image surface
of the objective lens 931. An opening device 933 is disposed on the
image surface, the opening device 933 having two confocal linear
openings 9331 and 9332 that correspond to the scatter images. The
image on the opening device 933 is enlarged and projected at lenses
980 and 985, with image sensors 991, 994 arranged at image
positions corresponding to the linear openings 9331 and 9332.
[0077] While an example of subjecting the illumination region 930
to high elevation angle TTL illumination and the illumination
region 929 to low elevation angle off-axis illumination has been
taken in the description of the present embodiment, the subjects of
the low elevation angle illumination and high elevation angle
illumination may be interchanged. In addition, an embodiment in
which polarizing and/or illuminating directions, illumination
wavelengths, and other conditions differ at the illumination
regions 929 and 930 would be probable.
Fifth Embodiment
[0078] The examples of detecting a plurality of images of different
detection conditions by splitting one optical path into a plurality
of paths have been shown in the detection systems described in FIG.
1 of the first embodiment, FIG. 8 of the third embodiment, and FIG.
9 of the fourth embodiment. The embodiment described below relates
to detecting images of multiple sets of conditions without
splitting an optical path of a detection system. FIG. 10 shows an
optical system configuration that uses high elevation angle TTL
illumination. Although layout of optical parts in this optical
system is very similar to the configuration described in FIG. 3,
the optical system differs in that it includes image sensors 1091
to 1094 and optical filtering devices 1095 to 1098 arranged in
immediate front thereof (low elevation angle illumination of a
wafer 1 uses the same method and same configuration as those
described in the first and second embodiments, so description of
the low elevation angle illumination is omitted). In immediate
front of the image sensors 1091 to 1094, the optical filtering
devices 1095 to 1098 each different in filtering characteristics
are arranged for the image sensor 1091 to 1094 to detect four kinds
of images that suit the filtering characteristics.
[0079] FIG. 11 shows a polarizing-based filter array as an example
of the optical filtering devices 1095 to 1098. The filter array is
composed of an optical filtering device 1095, which uses a
polarized-light transmitting axis parallel to a Y-direction, a
device 1096, which forms an angle of 45.degree. from the
Y-direction, a device 1097, which forms an angle of 90.degree. from
the Y-direction, and a device 1098, which forms an angle of
135.degree. from the Y-direction. This composition makes it
possible to detect the four kinds of images different in analyzing
conditions, in the same space on the wafer 1, and to perform
comparative inspections with different polarization states of
normal-pattern-scattered light and defect-scattered light as
feature quantities.
[0080] A general image sensor is formed on a silicon substrate.
However, a mode in which four sensors are arranged on one substrate
and four images are output would be probable as a modification of
the present invention. A photonic crystal with fine patterns
laminated thereupon, a wire grid, a dielectric multilayer film
structure, or the like are applicable as each optical filtering
device.
[0081] A flow of processing conducted by an image-processing unit
10100 is shown in FIG. 12. The image-processing unit 10100
processes the images that have been detected by the image sensors
1091-1094, and determines whether the images contain defects.
[0082] The image detected by the image sensor 1091 is subjected to
conversion of brightness, such as .gamma.-correction, in a
gradation conversion unit 1201a. One of two images by the
conversion is sent to a position-matching unit 1205a, and the other
is sent to a memory 1203a. The position-matching unit 1205a matches
a position of the image which has already been received directly
from the conversion unit 1201a and a position of the image stored
within the memory 1203a until does arrive the same predesigned
pattern (e.g., image of an adjacent die) as that of the directly
received image, and received from the memory 1203a. A comparator
1207a creates a differential image and the like from the two
position-matched images by conducting comparisons, and then
calculates feature quantities of the differential image as results
of the comparisons. Next, a defect-determining unit 1215 uses the
feature quantities (such as a maximum contrast level and area) of
this differential image to determine whether the image contains
defects.
[0083] Substantially the same processes as those described above
are also conducted upon each of the images detected by the image
sensors 1092, 1093, and 1094. Additionally, results of these image
comparisons are sent to a position-matching unit 1210. Then, the
position-matching unit matches the positions of the four images
different in illumination elevation, detection elevation, and
analyzing conditions. After comparison of feature quantities among
the images of the different optical conditions, these feature
quantities are sent to the defect-determining unit 1215, for defect
determination. Therefore, the defect-determining unit 1215 conducts
determinations using the five kinds of feature quantities in all.
If any one of the five sets of determination results indicate that
the corresponding image is defective, the feature quantities of the
remaining four kinds are also sent to a classification unit 1217.
The classification unit 1217 classifies detected defects by kinds
(e.g., foreign matter, etching residues, or scratches) or as dummy
defects (such as non-uniformity in brightness of an oxide film,
roughness of the pattern, grains, or other factors not critical or
fatal to the semiconductor device). After classifying, the
classification unit 1217 outputs coordinates, classifying results,
feature quantities, and others of the defects.
[0084] An example of feature quantities calculated from the four
images which were detected using the different optical conditions
is shown in FIG. 13. The optical system configuration that was
adopted in this example is that shown in FIG. 10, and the four
images in this example were detected using a polarized-light
transmitting axis parallel to a Y-direction, and polarized-light
transmitting axes angled at 45.degree., 90.degree., and 135.degree.
to the Y-direction. A polarization state 400 is calculated, pixel
by pixel on the same space of the wafer, from each of the images.
Indicators of the polarization state would be an azimuth a of
elliptical polarization, and an ellipticity ("b/a") calculated in
terms of a ratio between a major axis "a" and minor axis "b" of the
ellipse. A defect-determining technique using these feature
quantities is shown in FIG. 14. This figure shows a scatter diagram
with the azimuth plotted as a on a horizontal axis and the
ellipticity plotted as p on a vertical axis, the diagram being a
schematic of data plots having the same coordinates in design
pattern. Normal patterns are distributed in regions that resemble
in both azimuth .alpha. and the ellipticity .beta.. Points A, B,
and C that overstep the distribution are determined as defect
candidates.
[0085] While various combinations are possible for each of the
configurations, functions, and image-processing details shown and
described in the above embodiments, these combinations obviously
fall under the scope of the present invention.
INDUSTRIAL APPLICABILITY
[0086] The present invention can be applied to methods, and related
devices, for inspecting defects, foreign matter, and other unwanted
substances present on fine patterns formed on samples through a
thin-film process in semiconductor-manufacturing processes and
flat-panel display manufacturing processes.
DESCRIPTION OF REFERENCE NUMERALS
[0087] 1 . . . Wafer, 3 . . . XYZ.theta. stage, 4 . . .
Illumination system, 5 . . . Laser, 6 . . . XYZ.theta. stage, 7 . .
. Electro-optic element, 10 . . . Beam expander, 15 . . . Rotatable
half wavelength plate, 17 . . . Rotatable quarter wavelength plate,
22 . . . off-axis illumination light, 26 . . . TTL illumination
light, 27 . . . Beam splitter, 29 . . . Illumination region, 30 . .
. Illumination region, 31 . . . Objective lens, 32 . . . Mirror, 33
. . . Opened device for confocal detection, 40a . . . Opened device
of checkered pattern, 62 . . . Pupil, 80 . . . Spatial modulator,
90 . . . Image sensor, 100 . . . Image-processing unit, 110 . . .
Operating unit, 120 . . . Mechanism control unit, 130 . . . Height
detection unit, 400 . . . Polarization state
* * * * *